Despite a highly variable course, Alzheimer's disease and other dementias have a devastating impact on patients and their families. Throughout the inevitable progression of these diseases, family physicians can help patients and their caregivers by offering emotional support and providing resources for coping with difficult behaviors. Pharmacologic therapy is another option.

This issue of American Family Physician contains a summary of the recommendations of a joint American Academy of Family Physicians (AAFP) and American College of Physicians (ACP) guideline panel on medications for dementia.¹ The full guideline appears in Annals of Internal Medicine.² The guideline is based on a systematic review conducted by the Evidence-based Practice Center at McMaster University.³ It addresses five medications (donepezil [Aricept], galantamine [Razadyne], rivastigmine [Exelon], tacrine [Cognex], and memantine [Namenda]); it does not address other medications used for the treatment of neuropsychiatric symptoms (e.g., aggression, agitation), which are often associated with dementia. No medications are currently approved for treatment of neuropsychiatric symptoms.⁴

Physicians should base the decision to initiate pharmacologic therapy on an individualized assessment of potential risks and benefits. Four of the agents considered are cholinesterase inhibitors. Excluding tacrine, which has a worse side-effect profile, there were no significant differences among rivastigmine, galantamine, and donepezil.^2,3 The other agent considered, memantine, acts to limit stimulation of the glutamate system. Its effectiveness has been tested primarily in patients with moderate to severe dementia, and it is approved for those patients. Patients taking cholinesterase inhibitors and memantine have statistically significant improvements in cognitive function, but on average, the benefits are not clinically significant for cognition and are only modest for global assessments.² Some patients do achieve clinically important benefits; however, because of study variation in instruments and reporting, making a summary estimate of the percentage of patients who have a clinically important response is difficult.² Similarly, evidence is lacking in regard to the effect therapy has on important patient-oriented outcomes, such as institutionalization. The guideline calls for further research to address additional important questions, such as the effectiveness of combination therapy and the duration of treatment in those who respond.

A guideline from the National Institute for Health and Clinical Excellence (NICE) in Great Britain, published in November 2006, reached similar conclusions about medication effectiveness.⁵ In addition, there was no solid evidence that the currently available medications increase quality of life, delay institutionalization, or treat the neuropsychiatric symptoms in patients with dementia. NICE recommends consideration of cholinesterase inhibitors only in patients with moderate dementia, and does not recommend the use of memantine outside research studies.

Some may argue that, in the face of limited benefits, the social cost of using these expensive medications is not justified. Consideration of the cost of therapy, which is not addressed in the AAFP/ACP guideline, undoubtedly influenced the NICE recommendations. However, in patients with mild to moderate dementia, we believe the real problem for most physicians is not deciding whether to start these medications, but rather clearly defining the therapeutic goals. In our experience, most patients and families are willing to try anything that might help. Moreover, some patients do benefit. Unfortunately, we cannot predict which patients will respond and which will not.

Aside from managing intolerable side effects, the real question for family physicians is when, if ever, to discontinue these medications. What is a reasonable trial of therapy? If effective, what is the appropriate duration of therapy? Our experience is that once these medications are started, they are rarely discontinued, even when the patient continues to deteriorate. The AAFP/ACP dementia guideline panel could not answer these questions because of a lack of evidence, but they did state that a three-month trial is acceptable. They also emphasized that if slowing decline is no longer a goal (e.g., in advanced dementia), there is no reason to continue these medications.

So, how should family physicians treat dementia in light of this new guideline? It seems reasonable to discuss the risks and benefits of a trial of one of the currently available cholinesterase inhibitors to all persons diagnosed with mild to moderate Alzheimer's disease or vascular dementia. Use of memantine may also be suitable in patients with moderate cognitive impairment. The patient's status should be closely monitored for the effectiveness of the treatment. In addition to the physician's own assessment of the patient, the observations of the patient and family should be included in the evaluation of treatment effectiveness. A standardized measure of cognitive function, such as the Mini-Mental State Examination,⁶ and a measure of physical function should also be obtained. It would seem reasonable to continue medication if documented benefit results, including a halt to cognitive or functional decline. If the patient continues to deteriorate, stopping therapy would be acceptable.